Our voice library is brimming with wonderful and fantastical voices, and our pitch and pace controls already give you a lot of control over how they sound. But when you want something specific, that’s where Voice Lab shines. It lets you blend voices from our library to create a unique AI Voice for your character. Best of all, these voices can be saved to use later, offer the same controls as any other voice on our platform and remain consistent in accent and performance.
Today we’re going to show you how to create a custom, consistent AI Voice for your character with Voice Lab, without using prompts, for a more controllable output. We're going to use our concept character called Chaz. He’s been described as impish and a bit of a scamp. Start by selecting "blend a custom voice" in Voice Lab.
Choosing a Performer for the Custom AI Voice.
Voices in Voice Lab is currently made up of 2 components, the performer and the voice blend. The performer drives the character's accent and speech cadence. Our custom AI Voice will also inherit the various styles available to the performer. The voice blend lets us control the tone and timbre of the voice.
Lets start with the performer side. We’ve think that a child-like British cockney accent would work really well for Chaz. Bazur fits the bill nicely from a performance and accent perspective. He's typically cast as a Constable or a Watch Captain, but his performance will be perfect with a little help from Voice Lab.
And here's what he sounds like.
Adjusting the Speech Rate
This accent is just what we’re looking for, but I think the performance is a bit too slow for Chaz. So, let's crank the pace or "speech rate" up by about 50%. Replica's AI Voices use Neural Pace control to essentially make the AI models speak faster, rather than simply speeding up the audio after it has been generated.
Designing the AI Voice Blend
Now it's time to craft the sound of the voice. Given the character is imp and a bit child-like, I’m looking for a sound that reflects that. Let's start by adding one of our younger voices to the voice blend. Pumpkin is one of our younger characters, and sounds a bit “bratty”.
By default, Voice Lab will try to automatically identify the appropriate pitch for the voices in the blend, but this can be overridden. In this case, I think its a bit too high for Chaz. We'll change the pitch from Auto, to +8 semitones. Similar to our pace control, this is a AI or Neural pitch control, so the results can sound quite natural.
Chaz is starting to sound pretty great, but perhaps Pumpkin is a little too harsh. Let's include another voice to round things out a bit. Pink Dragon is excellent for cutesy high pitched characters. Voice Lab can blend up to 5 voices seamlessly, giving a huge amount of creative control. We'll create a mix of 50% Pumpkin and 50% Pink Dragon. By creating our custom voices in this way, we ensure that we can accurately measure the contribution of each of our Voice Actors into the finished product.
Using your Custom Voice for Text to Speech
We can now save our character as Chaz, and start using him to generate speech in Voice Director. Chaz will inherit the styles of the performer, is capable of Text to Speech, Speech to Speech with Voice Changer and Multi-language.
Here's Chaz!
We've come a long way, when you compare it Bazur straight from the library.
Next Steps
In a world where unique characters bring games to life, having the right voice can make all the difference. With our AI voice generator, creating custom voices for your game has never been easier. Whether you need a gritty warrior or a quirky sidekick, our tools let you design voices that fit perfectly with your characters. No need for complex audio setups—our intuitive interface makes it simple for anyone to dive in and craft distinct, high-quality voices. Empower your game’s storytelling with voices as unique as your characters, and elevate player experiences with seamless, easy-to-use customization. Get started making custom voices with Replica Studios today.
